Sleep aids in the consolidation of memory and also plays a key role in regulating emotions. Extinction is a form of emotional memory that promotes emotion regulation by inhibiting the expression of conditioned fear. Using a validated 2-day human protocol, we propose to compare extinction memory between persons with Primary Insomnia and normally sleeping controls to test the main hypothesis that those with insomnia will display poorer extinction memory. Conditioned fear will be established to two colors as conditioned stimuli (CS+) using a mild electric shock as an unconditioned stimulus (US) and skin conductance response (SCR) as the conditioned response (CR). A third color never paired with a shock (CS-) will confirm differential conditioning. Fear of one color (CS+E) will then be extinguished by its repeated presentation without the US. After 24 h, memory for this extinction learning as well as the conditioned fear memory for the unextinguished CS+ (CS+U) will be tested. The accompanying neural responses will be measured at each stage of this protocol (fear conditioning, extinction learning, extinction recall) using functional magnetic resonance imaging (fMRI). Sleep physiology will be measured using ambulatory polysomnography (PSG) on an acclimation night, a baseline night and on the night that will intervene between the day when conditioning and extinction are learned and when they are recalled 24 h later. Correlations of extinction memory performance with sleep physiology will be examined among sleep quality measures (e.g., sleep efficiency) as well as sleep architectural measures (e.g., sleep stage percentages) and measures specific to rapid eye movement (REM) sleep (e.g. REM continuity). It is hypothesized that extinction memory (the degree to which SCR to the CS+E remains suppressed after 24 h) will be poorer in insomniacs. Additionally, it is hypothesized that the degree to which extinction recall generalizes (i.e., the degree to which SCR to the CS+U as well as to the CS+E is reduced) will be greater in controls versus insomniacs. It is also hypothesized that, in insomniacs versus controls during extinction recall, the neural structures whose activity has been associated with extinction learning and memory (ventromedial prefrontal cortex and hippocampus) will be hypoactivated and those associated with the expression of conditioned fear (amygdala and dorsal anterior cingulate cortex) will be hyperactivated. Lastly, it is hypothesized that, in insomniacs, extinction memory will correlate positively with sleep quality as well as with the integrity of sleep architecture, especially during REM sleep, across the night intervening between extinction learning and extinction recall.